收稿日期:2022-08-12.*通信作者:胡丹峰E-mail:hdf78@suda.edu.cn光电技术及应用DOI:10.16818/j.issn1001-5868.2022081201基于双摄像头的心率测量系统的设计与改进徐杰,胡丹峰*,张林,王加俊(苏州大学电子信息学院,江苏苏州215006)摘要:针对在外部干扰或昏暗环境下利用图像光电容积描记(imagePhotoplethysmography,iPPG)技术进行心率测量时准确度较差,提出了一种自适应心率提取算法,并在嵌入式硬件平台上进行了验证。算法根据图像中人脸与背景区域的色度关系来识别不同的场景并启动合适的摄像头进行图像采集及自适应映射,接着对提取出的信号进行滤波,在信号质量评估后输出结果。上述方法在Zynq平台上进行了验证,使用双摄像头实现实时心率测量,并对结果进行可视化输出。实验结果表明:优化后的算法在光照及运动的双重干扰下的测量误差从3.36BPM降至2.78BPM,准确率提升了17.3%。另外,所设计的系统能够实现在极端黑暗条件下的心率采集,平均误差约为2.39BPM。关键词:图像光电容积描记法;心率测量;Zynq;抗干扰;信号处理中图分类号:TP391.41文章编号:1001-5868(2023)01-0115-07DesignandImprovementofHeartRateMeasuringSystemBasedonDualCameraXUJie,HUDanfeng,ZHANGLin,WANGJiajun(SchoolofElectronicandInformationEngineering,SoochowUniversity,Suzhou215006,CHN)Abstract:InordertosolvetheproblemthattheaccuracyofheartratemeasurementispoorunderexternalinterferenceordimenvironmentbyimagePhotoplethysmography(iPPG),anadaptiveheartrateextractionalgorithmwaspresentedandwasverifiedonanembeddedhardwareplatform.Differentscenescouldbeidentifiedbythisalgorithmaccordingtothechromaticityrelationshipbetweenfacesandthebackgroundareasintheimage.Subsequently,theappropriatecamerawaslaunchedforimageacquisitionandadaptivemapping.Finally,theextractedsignalwasfilteredandtheresultwasoutputafterthesignalqualityevaluation.ThealgorithmwasimplementedontheembeddedZynqsystemwithdualcamerasforreal-timeheatratemeasurement,andtheresultswerevisualized.Experimentalresultsshowthatthemeasurementerrordecreasesfrom3.36BPMto2.78BPMundertheinterferenceoflightandmotion,andtheaccuracyisimprovedby17.3%.Inaddition,thedesignedsystemcanalsoachieveheartrateacquisitionunderextremedarkconditions,withanaverageerror...
